CELLULOSE PROCESSING WITH CHLORIDE-BASED IONIC LIQUIDS

Transcription

1 CELLULOSE PROCESSING WITH CHLORIDE-BASED IONIC LIQUIDS Gino Bentivoglio 1,2, Thomas Röder 3, Mario Fasching 1, Mario Buchberger 3, Herwig Schottenberger 2, and Herbert Sixta 3. 1 Kompetenzzentrum Holz GmbH, St.-Peter-Str. 25, 4021 Linz, Austria; Phone: ; Fax: ; [email protected] 2 Faculty of Chemistry and Pharmacy, University of Innsbruck, Innrain 52a, 6020 Innsbruck, Austria 3 Lenzing AG, Department of Pulp Research, Werkstrasse 1, 4860 Lenzing, Austria Ionic liquids (ILs, salts with a melting point below 100 C) are discussed as solvents for cellulose with a potential for industrial applications. Several chloride containing ILs have been tested for their cellulose dissolving properties. Partly, strong cellulose degradation was observed, but could be prevented in some cases by addition of stabilisers. Cellulose degradation was compared for five chloride ILs. For three solvents, 1-butyl-3-methylimidazolium chloride, 1-allyl-3-methylimidazolium chloride and 1,3-diallylimidazolium chloride the temperature effect on degradation was studied. Fibres could be obtained by spinning the IL solutions into water; fibre characteristics are presented. The experimental cellulose spinning process with chloride containing ILs is compared to the well-known NMMO-based Lyocell process. Keywords: cellulose, ionic liquids, degradation, fibres, Lyocell process Introduction Ionic liquids [1a,b] Liquids consisting only of ions are called ionic liquids (ILs). In the broader sense, this term includes all kind of salt melts, like sodium chloride at temperatures above its melting point of 800 C. Today, the term ionic liquid refers particularly to salts with a melting point below 100 C. Salts with a melting point below 25 C are called room-temperature ionic liquids (RTILs). Usually, ionic liquids consist of a bulky, asymmetric organic cation, like 1-alkyl-3- methylimidazolium, 1-alkylpyridinium, 1- methyl-1-alkylpyrrolidinium or ammonium ions. A wide range of anions is employed, from simple halides which inflect high melting points, to inorganic anions such as tetrafluoroborate and hexafluorophosphate and to large organic anions like bis(trifluorosulfonyl)amide, triflate or tosylate. The notable characteristics of ionic liquids are their non-measurable vapor pressure, thermal stability, wide liquid range, electric conductivity and solvating properties for diverse kinds of materials. Another important feature of ionic liquids is their designability: miscibility with water or organic solvents can be tuned through sidechain lengths on the cation and choice of anion. Furthermore, their properties can be varied by introduction of functional groups. Because of their specific properties, ionic liquids have found to be useful in many fields, like as reaction media in organic synthesis or electrolytes for electrochemical applications. In addition, their non-volatility results in low impact on the environment and human health advantaging them in comparison to conventional organic solvents, and they 154

2 are recognized as green solvents. However, this term is misleading, as many ILs show aquatic toxicity [2] and spilling into waterways should be avoided. Ionic liquids as solvents and reaction media for biomacromolecules The search for suitable cellulose solvents was and is still of great importance in cellulose research, as shaping of cellulose can only be achieved by dissolution and regeneration. The fact that cellulose is earth s most abundant renewable biomacromolecular resource will encourage further development of cellulose-based materials, whose production will require suited solvents. Today, only one solvent system with structural analogies to ionic liquids is used commercially on large scale, namely N- Figure 1. NMMOmethylmorpholine-N- monohydrate. oxide monohydrate (NMMO, Figure 1). The cellulose is dissolved without derivatisation and subsequently spun (Lyocell-process). This system has already been extensively described in the literature. [3] As many ILs, NMMO possesses a quaternized nitrogen atom. Contrarily to ionic liquids, introduction of alkyl substituents into the heterocyclic ring of NMMO decreases the solubility of cellulose. The application of low-melting quaternised nitrogen bases as solvents for cellulose was first proposed by Graenacher in [4] In his patent, he describes the preparation of cellulose solutions in benzyl pyridinium chloride in the presence of nitrogen bases, possibly derivatisation and subsequent regeneration of the cellulose (e.g. as threads or films) by precipitation in water or alcohols. However, these results were thought to be of little practical value and the invention found no application. In 2002, the term ionic liquid was first mentioned in Figure 2. BMIM-Cl. the context of cellulose. Rogers & coworkers investigated the cellulose dissolving ability of several dialkylimidazolium-based ILs. [5] With 1- butyl-3-methylimidazolium chloride (BMIM-Cl, Figure 2), using a pulp with polymerisation degree of ~ 1000, cellulose solutions up to 10 % could be obtained. The dissolved cellulose was subsequently precipitated in water or organic solvents like ethanol or acetone. Water in concentrations above 1 % prevents the cellulose from being dissolved. By microwave irradiation solutions with 25 % cellulose were prepared. This high solubility under microwave irradiation suggests that decomposition takes place; unfortunately, no information regarding the degradation of the cellulose was given by the authors. Since then, a number of cellulose dissolving ionic liquids based on imidazolium cations have been found. In 1-allyl-3-methylimidazolium chloride (AMIM-Cl) cellulose concentrations of 14.5 % (DP of dissolving pulp of ~ 650) and 8 % (DP ~ 1600) could be achieved. [6] 1-Allyl-3-butylimidazolium chloride (ABIM-Cl) and 1,3-diallylimidazolium chloride (AAIM-Cl) are solvents for cellulose as well. [1a] Among the multitude of tested ILs, especially the chlorides have shown outstanding cellulose dissolving abilities. The exact dissolution mechanism is still unknown; however, it is assumed that the unhydratised chloride anions present in the ionic liquid are able to disrupt the strong intra- and intermolecular hydrogen bonds between the cellulose chains. This assumption is supported by the fact that the dissolving ability increases with the chloride ion concentration. [7] A number of publications deal with the derivatisation of cellulose in ionic liquids. Cellulose acetate with a substitution degree of can be produced in a homogeneous reaction with AMIM-Cl as solvent. [8] In ionic liquids, acylations with various acid chlorides and 155

3 carbanilations are possible under mild conditions without the need of a catalyst and in short reaction times. [9] Cellulose ethers can be synthesised in ionic liquids with slight excess of reagents under mild conditions in a water free environment. [10] Besides cellulose, other biopolymers can be dissolved in ionic liquids, as it is known from NMMO. Lignocellulosic material like straw or wood in the form of chips or sawdust can be dissolved in BMIM-Cl under microwave irradiation. This opens an easy way to isolate individual wood components, for instance by fractional precipitation. [11] Wool (keratin) could be dissolved in BMIM-Cl with concentrations up to 11 % and regenerated by precipitation into water or alcohols. By adding a solution of cellulose in BMIM-Cl prior to precipitation, wool-cellulose composites in form of fibres or membranes were obtained. [12] Ionic liquids have also proven to be good solvents for silk. For instance, the saturation concentration of silk fibroin in 1-ethyl-3- methylimidazolium chloride (EMIM-Cl) at 100 C is about 23 %. [13] By spinning 10 % solutions into methanol regenerated silk fibres were obtained. [14] Ionic liquids as solvents in the spinning process A method for the production of cellulosic moulds like fibres or films from ionic liquids was described by the Thuringian Institute of Textile and Plastics Research. [15] After dispersion in water, the moist cellulose is mixed with aqueous BMIM-Cl solution under addition of stabilizers like sodium hydroxide and propyl gallate. Under shear strain, temperature, and vacuum the suspension is transformed into a homogeneous, nearly water free dope. By passing through a spinneret and an air gap, the solution is shaped into fibres or foils. The cellulose is regenerated by precipitation in an aqueous spinning bath. To regenerate the solvent, the spinning bath is treated with alkaline hydrogen peroxide solution, metal ions are removed with the aid of an ion exchanger and the water is finally removed by distillation. According to the Bisfa-Definition, this process is a Lyocell-process. [16] Fibres obtained by this procedure are very similar to Lyocell fibres obtained by the NMMOprocess, due to the comparable dissolution step, the similar solution structure, and the same regeneration conditions. The so called working capacity (tenacity* elongation) in dry conditions is the same. The tendency to fibrillation is comparable, too. Other patents concerning the production of cellulose fibres from ionic liquids have been applied by Chinese groups. [17, 18] Experimental Ionic liquids Ionic liquids were synthesised according to known procedures. [1a, 19, 20, 21] Synthesis of 1-Allyl-2-methylpyridinium chloride To 20.0 ml of 2-picoline (18.9 g, 0.20 mol, 1 eq.) was added an excess of allyl chloride (20.0 ml, 18.6 g, 0.24 mol, 1.2 eq.). The reaction mixture was refluxed for 48 h. After 1 h, a dark brown, heavier phase began to separate. TLC analysis (Merck silica gel, ethyl acetate) showed traces of starting material. After refluxing for additional 24 h, excess allyl chloride was removed by means of an oil pump leaving behind a brown solid. The crude product was washed with 50 ml of diethyl ether and finally dried on a high vacuum line giving 17.4 g of 1-allyl-2- methylpyridinium chloride (brown powder, 51 % of theory). Analytical data: 1 H NMR (CDCl 3 ): δ 2.93 (3H, s), 5.10 (1H, d, J 17.2 Hz), 5.35 (1H, d, J 10.6 Hz), 5.66 (2H, d, J 5.6 Hz), 6.00 (1H, m), 7.92 (1H, t, J 6.8 Hz), 8.00 (1H, d, J 7.9 Hz), 8.41 (1H, t, J 7.6 Hz), 9.70 ppm (1H, d, J 5.9). 13 C NMR (CDCl 3 ): δ 20.5, 60.0, 120.9, 126.2, 130.0, 130.1, 145.5, 146.9, ppm. IR (neat, ATR): 3009, 2921, 2438, 1622, 1573, 1503, 1478, 1455, 1421, 156

4 1296, 1158, 1141, 1053, 1004, 930, 829, 794, 770, 710, 663 cm -1. Mp.: C. Dissolution experiments Chloride based ionic liquids have been tested in the face of their cellulosedissolving ability. For this, the preparation of 3 % solutions of cellulose in the ionic liquid was attempted. Beech sulfite pulp was hackled in a kitchen blender, mixed with the ionic liquid, and stirred magnetically in a teflon coated reaction vessel at 100 C for 2 h. If no dissolution took place under these conditions, stirring was continued for another 2 h at 110 C. Then, the solution was examined under a light microscope to reveal undissolved fibres. To determine the degradation, 3 % cellulose solutions were prepared and the dissolved cellulose reconstituted by contacting with water. The molecular mass distribution of the reconstituted cellulose samples was determined by gel permeation chromatography using DMA/LiCl as eluent. The method was described in detail earlier. [22] Spinning experiments Spinning dopes with concentrations over 10 % were prepared in a vertical kneader. For this, the dry pulp was added to the solvent under optional addition of propyl gallate and/or sodium hydroxide as stabilisers and transformed into a homogenous solution under shear strain, temperature, and vacuum. Results and discussion [23] Besides the well-known cellulose solvents AMIM-Cl and BMIM-Cl, new solvents with imidazolium- and methylpyridiniumcations were found (Table 1). The assumption, that a N-O-bond in the cation would be favorable for cellulose dissolution (structural analogy to N- methylmorpholine-n-oxide), proved to be false. Cellulose was insoluble in the tested N-alkyloxyimidazolium salts. Ionic liquids functionalised with hydroxyl groups didn t dissolve cellulose as well. Table 1. Solubility of 3% cellulose in ILs (chlorides). Ionic liquid 1-Butyl-3-methylimidazolium chloride (BMIM-Cl) + 1-Allyl-3-methylimidazolium chloride (AMIM-Cl) + 1-Allyl-3-butylimidazolium chloride (ABIM-Cl) + 1,3-Diallylimidazolium chloride (AAIM-Cl) + 1-Allyl-2-methylpyridinium chloride (A2Pic-Cl) + 1-Butyl-2,3-dimethylimidazolium chloride + 1-Allyl-3-propargylimidazolium chloride reacts 1-Allyloxy-3-methylimidazolium chloride 1-Allyl-3-Hydroxyethylimidazolium chloride 1-Methyl-3-Hydroxyethylimidazolium chloride Degradation behaviour The degradation of the dissolved polymer is of great importance to determine the suitability of a solvent system. Rogers & coworkers describe the dissolution of 25 % cellulose in BMIM-Cl with the aid of microwave irradiation. By means of a vertical kneader, 25 % solutions are feasible, too. However, the cellulose is subjected to strong degradation (Figure 7). These results point out the need of a stabiliser to prevent cellulose degradation. With conventional dissolving pulps, cellulose concentrations in the spinning dope of more than 15 % are not within reach due to the high viscosity of such solutions. Higher temperatures lead to stronger degradation; therefore it has no advantage to the NMMO-process. It was found, that the stability of the cellulose depends clearly on the used cation (Table 2). All of the tested solvents based on chloride showed conspicuous degradation of the cellulose at 100 C. The degradation was exceptionally strong in AMIM-Cl and A2Pic-Cl, whereas cellulose regenerated from ABIM-Cl and AAIM-Cl showed relatively high molar masses. Table 2. Molecular masses of cellulose after dissolution in various ILs. Solvent M n (x 10 3 ) M w (x 10 3 ) pulp: Euca-PHK (1) BMIM-Cl AMIM-Cl A2Pic-Cl AAIM-Cl ABIM-Cl

5 1 0 ABIM-Cl AAIM-Cl BMIM-Cl A2Pic-Cl AMIM-Cl Euca-PHK (1) Figure 3. Molecular mass distribution of cellulose after regeneration from ILs. The effect of the temperature on degradation was examined on the wellknown solvents BMIM-Cl and AMIM-Cl and on the new solvent AAIM-Cl. At the same time, the effect of propyl gallate, a stabiliser with antioxidative action used in the NMMO-process, was investigated. BMIM-Cl: As expected, the extent of degradation increased with the temperature. Above 90 C, drastic degradation occurred. Propyl gallate had no stabilising effect under these conditions. Table 3. Molecular masses in BMIM-Cl at various temperatures, with/without propyl gallate. Not stabilised propyl gallate M n M w M n M w ( x 1000 ) ( x 1000 ) ( x 1000 ) ( x 1000 ) 80 C C C ,2 BMIM-Cl 1,0 0,8 0,6 0,4 0,2 0,0 no stab. 80 C no stab. 90 C no stab. 100 C stab. 80 C stab. 90 C stab. 100 C Euca-PHK (1) Figure 4. BMIM-Cl at various temperatures, with/without propyl gallate. AMIM-Cl: A distinct temperature effect could be observed. The degradation was pronounced even at 80 C. Propyl gallate showed no significant influence on the degradation behaviour. Table 4. Molecular masses in AMIM-Cl at various temperatures, with/without propyl gallate. Not stabilised Propyl gallate M n M w M n M w ( x 1000 ) ( x 1000 ) ( x 1000 ) ( x 1000 ) 80 C C C ,6 AMIM-Cl 1,4 1,2 1,0 0,8 0,6 0,4 0,2 0,0 Euca-PHK (1) 80 C, no stab. 90 C, no stab. 100 C, no stab. 80 C, stab. 90 C, stab. 100 C, stab. Figure 5. AMIM-Cl at various temperatures, with/without propyl gallate. AAIM-Cl: This solvent showed temperature dependent degradation, too. However, compared to the other two solvents, the degradation occurred to a much lesser extent. Propyl gallate had no stabilizing effect. 1,2 1,0 0,8 0,6 0,4 0,2 0,0 AAIM-Cl Euca-PHK (1) 80 C, no stab. 90 C, no stab. 100 C, no stab. 80 C, stab. 90 C, stab. 100 C, stab. Figure 6. AAIM-Cl at various temperatures, with/without propyl gallate. The experiments showed clearly, that the substituents on the cation have an effect on the stability of the dissolved cellulose. Propyl gallate was useless under these conditions. 158

6 Table 5. AAIM-Cl at various temperatures, with/without propyl gallate. Not stabilised Propyl gallate M n M w M n M w ( x 1000 ) ( x 1000 ) ( x 1000 ) ( x 1000 ) 80 C C C Spinning experiments BMIM-Cl: The addition of propyl gallate in conjunction with sodium hydroxide has a stabilising effect. [15] Fibres could be spun from 11 % stabilised solutions. Table 6. Stabiliser effect on the molecular masses of cellulose in BMIM-Cl. M n [10 3 M w [10 3 g/mol] g/mol] Euca-PHK (2) stabilised not stabilised ,2 1,0 0,8 0,6 0,4 0,2 0,0 Euca-PHK (2) stab. n. stab. Log Molar Mass Figure 7. Molecular mass distribution of reconstituted cellulose out of stabilised 11 % solution (red) and not stabilised 25 % solution (green) in BMIM-Cl. 100 C, stirring time 3-4 h. AMIM-Cl: 15 % solutions could be obtained at 100 C within 2 h. However, the cellulose was strongly degraded by this procedure to an extent which affected the spinnability of the solution. By stirring for 5 h at a temperature of 70 C, 11 % solutions were obtained. Due to the lower temperature, the degradation remained on a reasonable scale and spinning was possible. Table 7. Temperature effect on molecular mass of cellulose in AMIM-Cl dopes. Mn (x1000) Mw (x1000) Euca-PHK (2) C C Euca-PHK solution (100 C) solution (70 C) Figure 8. Molecular mass distribution of reconstituted cellulose out of 15 % solution (red, 100 C, stirring time 2 h) and 11 % solution (blue, 70 C, stirring time 5 h) in AMIM-Cl. Fibre characteristics The spinning conditions in our experiments are similar to the spinning conditions in the NMMO-process. The resulting fibres show comparable characteristics, like a round profile, smooth surface, similar fibrillation behaviour and the same working capacity. The observed, somewhat higher tenacities are compensated by lower elongation values. Advantages of IL- Lyocell-fibres over NMMO-Lyocell-fibres could not be observed. From our point of view, there is no significant difference in the physical and chemical characteristics. Cellulose dissolution: Chloride containing ionic liquids compared to NMMO A comparison with the NMMO-process shows no improvement of the fibre characteristics, especially of the fibrillation tendency. The similar processes lead to similar fibre properties. ILs have a lower melting point than NMMO, which simplifies handling of the solvents. However, the required processing temperatures are the same in both cases. 159

7 Table 8. Comparison of fibre data for the solvents BMIM-Cl, AMIM-Cl, and NMMO Fibre Titre Tenac. (cond) Elong. (cond) Tenac. (wet) Elong. (wet) Working capacity (cond) dtex cn/tex % cn/tex % %cn/tex BMIM-Cl* BMIM-Cl* AMIM-Cl* TENCEL (NMMO)** TENCEL (NMMO)** * experimental fibre; **standard fibre Stabilisers are required in both cases. The preparation of the spinning dopes is more expensive than in the NMMO-process, because of the required nearly complete removal of water. An advantage is the thermal stability of the system, no autocatalytically initiated exothermic runaway-events were observed. The degradation of the cellulose is higher than in NMMO. Little is known about the reaction mechanisms, except that they differ from those in the NMMO-system. As for the solution process, the recovery of the solvent from the aqueous spinning bath is more energy-consuming because of the need for complete water removal. For industrial processes, the high corrosivity of chloride melts to steel may be of further concern. The potential toxicity of the ILs is also a disadvantage. For this reasons, we don t consider the tested chloride-based ILs to be an alternative to the NMMO system. Conclusions Chloride-based ionic liquids are suitable solvents for cellulose dissolution and for fibre spinning. The resulting fibres belong to the class of Lyocell-fibres and show comparable or same characteristics as fibres obtained from NMMO solutions. The advantages of ionic liquids, like their non-volatility, thermal stability, chemical modifiability, and low melting points are countervailed by their numerous disadvantages: the need for stabiliser use, potential (aquatic) toxicity, corrosivity, and a higher energy input for dope preparation and solvent recovery due to the required complete removal of water. In their textile quality, IL-fibres are virtually indistinguishable from conventional Lyocell-fibres on NMMObasis. Up to now, an industrial application of the tested IL-systems for the production of man-made cellulosic fibres is not useful. Firstly, none of these systems showed significant advantages in comparison to already used technologies; secondly, the recovery of the solvent is more expensive than in the NMMO process. Acknowledgement Financial support was provided by the Austrian government, the provinces of Lower Austria, Upper Austria and Carinthia as well as by the Lenzing AG. We also express our gratitude to the Johannes Kepler University, Linz, the University of Natural Resources and Applied Life Sciences, Vienna, and the Lenzing AG for their in kind contributions. 160

8 References [1] a) Laus, G.; Bentivoglio, G; Schottenberger, H; Kahlenberg, V.; Kopacka, H.; Röder, T.; Sixta, H.; Lenzinger Berichte, 2005, 84, 71-85; ref. cit.; b) Zhu, S.; Wu, Y.; Chen, Q.; Yu, Z.; Wang, C.; Jin, S.; Ding, Y.; Wu, G.; Green Chemistry, 2006, 8, [2] C. Pretti, C. Chiappe, D. Pieraccini, M. Gregori, F. Abramo, G. Monnia, L. Intorre; Green Chem., 2006, 8, [3] Rosenau, T.; Potthast, A.; Sixta, H.; Kosma, P. Prog. Polym. Sci. 2001, 26, [4] Graenacher, C. Cellulose Solution. U.S. Pat. 1,943,176, [5] Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. J. Am. Chem. Soc. 2002, 124, [6] Zhang, H.; Wu, J.; Zhang, J.; He, J. Macromolecules 2005, 38, [7] Remsing, R.; Swatloski, R.; Rogers, R.; Moyna, G.; Chem. Commun., 2006, [8] Wu, J.; Zhang, J.; He, J.; Ren, Q.; Guo, M. Biomacromolecules 2004, 5, [9] Barthel, S.; Heinze, T.; Green Chemistry, 2006, 8(3), [10] Myllymaeki, V.; Aksela, R.; Int. Pat. Appl. WO A (2005); Chem. Abstr. 143: [11] Myllymaeki, V.; Aksela, R.; Int. Pat. Appl. WO A (2005) Chem. Abstr. 142: [12] Xie, H.; Li, S.; Zhang, S. Green Chem. 2005, 7, [13] Phillips, D. M.; Drummy, L. F.; Conrady, D. G.; Fox, D. M.; Naik, R. R.; Stone, M. O.; Trulove, P. C.; De Long, H. C.; Mantz, R. A. J. Am. Chem. Soc. 2004, 126, [14] Phillips, D. M.; Drummy, L. F.; Naik, R. R.; De Long, H. C.; Fox, D. M.; Trulove; P. C.; Mantz, R. A. J. Mater. Chem. 2005, 15, [15] Michels, C.; Kosan, B.; Meister, F; Verfahren und Vorrichtung zur Herstellung von Formkörpern aus Cellulose. Int. Pat. WO , [16] Terminology% doc, aufgerufen am [17] Wang, H.; Liu, W.; Li, D.; Tu, X.; Zhao, T.; Yang, B.; Zhang, Y.; Yu, M.; Chinese patent CN A (2006) Chem.Abstr. 145: [18] Wang, H.; Liu, W.; Li, D.; Tu, X.; Zhao, T.; Yang, B.; Zhang, Y.; Yu, M.; Chinese patent CN A (2006) Chem.Abstr. 145: [19] Zhang, H.; Wu, J.; Zhang, J.; He, J. Macromolecules 2005, 38, [20] Mizumo, T.; Marwanta, E.; Matsumi, N.; Ohno, H. Chem. Lett. 2004, 33, [21] Andre, M.; Loidl, J.; Laus, G.; Schottenberger, H.; Bentivoglio, G.; Wurst, K.; Ongania, K.-H. Analytical Chemistry 2005, 77, [22] Schelosky, N.; Röder, T.; Baldinger, T. Das Papier 1999, 53, [23] Results presented in part as a poster at the 231st ACS National Meeting, March 26-30, 2006 Atlanta, GA. 161